I made an investment banking bot
What happens when a bot can do an investment banking analyst's job?
This morning while drinking coffee I made an investment banking analyst bot in ChatGPT. It took me all of five minutes. The bot isn’t going to get hired by Goldman Sachs any time soon, but it augurs a much different world.
Let’s dive in.
Here are the bot’s instructions:
Cash Flow Analyst is an expert tool for analyzing company cash flow statements, focusing on understanding sources and uses of cash and determining free cash flow. It delves into detailed analyses of cash flow statements, covering operating, investing, and financing activities. The tool calculates and interprets free cash flow, offering insights into financial health and potential for growth. It also evaluates key financial metrics, ratios, and trends for comprehensive performance assessment and future financial projections. Cash Flow Analyst communicates using technical financial language, catering to users with a solid understanding of financial terminologies and concepts.
This sounds impressive! How does it perform in real life? I gave it the following prompt: Review WeWork’s cash flow statement and give me an analysis of the company’s performance. Link: https://finance.yahoo.com/quote/WE/cash-flow?p=WE
The bot automatically fired up a browsing session in the background, and provided the following analysis:
Operating Cash Flow: WeWork has consistently reported negative operating cash flows, indicating a struggle to generate cash from its core business operations. The operating cash flow improved from -1,911,937 in 2020 to -728,000 in 2022, suggesting some operational improvements.
Investing Cash Flow: Negative investing cash flows across the years indicate significant investments, possible in capital expenditures or acquisitions. The sharp decrease from -4,775,520 in 2019 to -242,000 in 2022 could reflect a scaling back of these investments.
Financing Cash Flow: The company has had varied financing cash flows, with a significant influx in 2020 (2,337,971) and a reduction to 583,000 in 2022. This fluctuation could be due to varying levels of debt issuance and repayment or equity financing activities.
End Cash Position: Despite negative operating and investing cash flows, the end cash position has remained relatively stable due to financing activities. The end cash position was 245,000 in 2022, down from a peak of 934,999 in 2020.
Free Cash Flow: Free cash flow has been consistently negative, indicating that the company is consuming more cash than it is generating. The free cash flow improved from -3,977,065 in 2019 to -1,007,000 in 2022, but it remains a concern.
Capital Expenditure and Debt Activities: Capital expenditures have been decreasing, which might be part of a strategy to conserve cash. Issuance of debt has been a significant source of cash, especially in 2020 and 2022, but debt repayment activities have also been notable.
Overall, WeWork’s cash flow statement shows a company that is struggling to generate positive cash flow from operations, relying heavily on financing activities to maintain liquidity. The trend towards reduced negative cash flows and a stable end cash position might suggest some improvement in financial management, but the consistently negative free cash flow raises concerns about the long-term sustainability of the business model.
This is all very good, and it accords with my previous post about WeWork. You can check the bot out here.
I made the bot on the spur of the moment, as I had just opened up a new session in ChatGPT, and saw that I got access to its new GPT1 tool.
This bot is pretty simple. You can imagine a more robust one doing a lot of what an entry level investment banking analyst does. As the saying goes, “today’s AI is the worst you’ll ever use.” As I’ve said repeatedly, AI technology is rapidly improving, and these bots will become more powerful over time as abilities are developed, people learn best practices, etc.
The big investment banks will likely be exploring bots for their own internal use. So will many other companies which employ white collar knowledge workers.
So what does all of this mean for the future of white collar knowledge work? For the immediate future, probably not much. But the directional arrows of progress are evident. The question that has to be on everyone’s mind right now is whether a future version of this technology will take everyone’s job, or whether that future version of this tech will serve as a complement to, and productivity enhancer for, human labor.
Here, GPT stands for, confusingly, '“General Purpose Technology,” whereas the GPT in ChatGPT stands for “Generative Pre-trained Transformer.” No, I don’t understand why OpenAI has decided on such a confusing naming convention.